Reflection AI ex-employee says he was fired for using an AI agent and GPUs
teortaxesTex · x · 2026-11-08
A Reflection AI ex-employee published a detailed account of his termination: a 6 AM West Coast call (3 AM for him in Hawaii), an HR rep hired that same week reading out the reasons, his laptop locked five minutes later, and the timing right before an approved vacation for the ACL conference. The cited reasons: ignoring a prior "Level 3 warning," consuming other teams' GPU quota while using an AI agent to do his work, and colleagues spending too much time reviewing his code. He defends the agent as a workaround for the scheduler's lack of auto-resume for preempted experiments, claiming it let him scale dataset curation from tens of thousands to 100,000 entries in a week and 1 million over a weekend. Commenters note the disgruntled-employee bias but argue Reflection must respond, while benchmarking its models (Inkling 2S roughly pretraining-matched with DeepSeek V2, Beam near V3).
Related event: Ex-Employee Exposes Google-Style Bureaucracy at Reflection AI(4 posts)→
More from Companies & People
- Cisco logs $9.3B AI infra orders, buys three firms to push into agent security — shashib · 2026-11-08
- Sharp jab at OpenAI's Roon: mistaking 'abrasive self-importance' for 'lust for life' — teortaxesTex · 2026-11-08
- Asahi Linux ditches MIT for GPLv2 after LLM-rewritten code lifting incident — steipete · 2026-11-08
- Master's ROI top-100 ranking: Stanford +16.8%, worst schools near -13% — deedydas · 2026-11-08
- Reflection Exec Defends 100-PR-a-Day AI-Native Culture Against Fired Employee's Claims — teortaxesTex · 2026-11-08
- Stanford prof: AIxBio is enriched with grifters, integrity badly lacking — anshulkundaje · 2026-11-08